The popular framework of cortical entrainment postulates that speech comprehension crucially depends on the continuous alignment of low-frequency cortical oscillatory activity with the amplitude envelope of perceived acoustic speech signals. The evidence for cortical entrainment mostly stems from tightly controlled experimental paradigms focusing on repeated perception of isolated sentences that feature a very constant speaking rate. However, these kinds of decontextualised and extremely regular stimuli do not reflect natural speech as we encounter it in real life. We thus advance the view that naturalistic experimental paradigms, utilising spontaneously produced speech as stimuli and suitable frequency-domain methodological tools, should be used to address an important question that remains open: whether cortical entrainment is observed during speech perception and comprehension in real-life communicative situations. In addition, we discuss how the phenomenon currently labelled as cortical entrainment might be confounded by a regular repetition of evoked responses.